For Want of a Comma or Two

By Pete Cape, Global Knowledge Director

Here’s an interesting question from we received from a client: “How many moms of girls under 18 are there in the US?” A few commas in the sentence might have helped us work out exactly what was being asked. Who was under 18; the Mom or the girl? I read it one way, my colleague read it another.

So I worked it out both ways. Or at least tried to. Firstly: How many 15-17 year-old moms are there in the USA right now who have children who are girls?

The CDC (Centers for Disease Control and Prevention) in the US is a great source of statistics and they have data on teenage pregnancy. Normally, when it comes to demographic data I’ll take any fairly recent year and work from that. It generally moves slowly. Teen pregnancy however, has, thankfully, been dropping quite dramatically in the past few years. According to CDC it almost halved between 2017 and 2015 from around 41 per 1000 women 15-19 to 22 per 1000 in 2015. If this trend continues you can imagine it might be as low as 18/1000 this year.

Now the CDC data is still useful, and the “per 1000” is not the real problem. The problem is that the client is interested in 15-17 year-olds (under 18’s) and the CDC data relates to 15-19 year-olds. You have to ask yourself if there is an accelerating rate of pregnancy and birth among teens as they pass 17 and become 18 and then 19 years old. I feel the answer would be yes, but to what degree? This data is collected, but hard to find. Most often ‘Age Of First Birth’ data is presented as Mean Age, which is rising, as you might imagine knowing what we do now about teenage pregnancy. Poking around a bit more in the CDC I found enough numbers to estimate that in about 70% of cases of births to teenage Moms the Mom was 18 or 19 at the time.

Now our final consideration is time. We are not interested in births this year, or births last year or even in births at all, but in the “right now.” Someone who was 17 last year and had a baby last year is now 18 and still has a baby but should not be counted in our calculation.

We are only really interested in summing the green squares:

This of course requires us to know the distribution of ages of 15, 16 and 17 year-old Moms, this is not easily found so we need to make some reasonable estimate, so I took this:

15        5%

16        25%

17        70%

Which means of course that the vast majority of all younger “teen moms” from the past three years are excluded since they are now 18 or over.

Pulling all of the data together gave me a grand total of just shy of 78,000 moms who, in 2017, are 15, 16 or 17 years old. Some may have had more than one child in the period, some will have had twins, but we’d expect this to be small. So we assume they all have one child only. The last calculation we need to do is to work out how many of them are “moms of girls.” Every year there are slightly more boys born than girls, about 51% boys to 49% girls. For those who wonder why there are more women than men in the population? Simple; Men find more foolish ways of killing themselves before the age of 40, and women live longer generally!

So the answer to our question is that there are around 38,000 girls currently aged 15, 16 or 17 who are “moms of girls.” That’s roughly one in 100 15-17 year-old girls.

Secondly then: How many moms are there, right now, in the USA, who have girls who are aged 17 or under?

You can easily find out how many girls there are in the US under the age of 18 (it is 36,086,329 in 2017), and (almost) all of them have a living mom. The problem is that some of them are siblings of each other, so there are fewer moms than you first think.

From the Census Bureau’s American Community Survey for 2011, I found a distribution of households by number of children. The table comes in two varieties, one considering “own children” and the other considering “related children.” I took the “related children” table. This time we are going to ignore the fact that there are more boys than girls and assume 50:50 as this makes the math a lot easier to do! Further assuming no relationship between size of household and number of girls present, you can work out how many families with girls there are in the country.

So for each family with one child half of them have a girl, all of who have a mom. This is 7,951,317 moms.

For those with two children they have either zero, one or two girls. Since the probability of having a girl is independent and is 50%, we can work out how many families have what type of arrangement. All of them have moms (notice that I have to ignore re-combined families in all this …) and 75% of all two children households should have a girl in it.

This is all just simple combinations. Here are all the possible organisations of all two child households by birth order:

Boy Boy

Boy Girl

Girl Boy

Girl Girl

Three out of the four possible arrangements have a girl in it, so 75% of two child household have a girl in them. And so this is another 10,060,536 moms. On to three kids – 87.5% chance of having at least one girl, 4,751,315 moms. And the last category is “4+”…. which is a problem. If we assume the largest part of these are four child families they have a 93% chance of having a girl, some are five children households at 97% chance of having at least one girl. A few have six kids (98.5% chance of having a girl). Let’s call it 95% and then that’s 2,280,709 moms.

Total of 25,043,877 moms who have girls who are under 18. This is 18% of all women aged 15 or over, and it’s 23% of all women aged 15-65.

So the commas are important, as is asking for clarification when you’re asked to work out a number. Are these numbers right? No, but they are probably right enough to work out feasibility or rough incidence rates. And often that is all that is required.